Bill James Guide to Baseball Managers, The (49 page)

There had long been a feeling that relief aces were being overused. That was what John Thorn was responding to. He was saying, in essence, that most people may think that Mike Marshall was pitched to death in 1974, but then, people have been saying the same thing for sixty years. Each year there were more and more games
not
completed by the starting pitcher, and, since the reliever’s job was to pitch in any close game when the starter wasn’t available, each year there was more and more work to be done by the relief ace. There was no third clause to the definition of the relief ace’s job. Herman Franks wrote the third clause: The relief ace will work when the game is close, the starter is gone,
and it is a “save” situation
.

Within a very few years, this would become the standard definition of a closer’s job.

Now, it is very much an open question whether this is an optimal strategy. Does an effective relief ace have the greatest impact on the won-lost record of his team when he is used only to close out narrow
victories
, or might he not have equal impact, or even greater impact, if he was used as Elroy Face was used? I don’t know the answer to this question; I don’t think anybody else knows, either. There are no studies of the issue. There is no compelling logic. I have never heard any major league manager, executive, or ex-manager explain
why
using his reliever in this way is more effective than using him in some other way.

The manager who defined this practice, Herman Franks, was fired within a year, and never managed again.

Nonetheless, this practice is universally accepted by current major league managers. There is no exception. There is no manager who has continued to use his relief ace in the way that relievers were used from 1950 to 1978. There is something called the “Bullpen by Committee,” which means rotating the closer work among several pitchers, but even this, though often talked about, is rare in practice and is really nothing more than a transitional stage while the team attempts to identify the one pitcher who should be The Closer.

This is rather remarkable, isn’t it? Nobody studies the issue, nobody explains or defends the strategy in any meaningful way, and there is no dramatic example of a team which adopts this strategy and comes out of nowhere to win a pennant. And yet, within a few years, every major league manager adopts the strategy. How can that happen?

Our language drives our thought.

Jerome Holtzman developed the concept of saves for a reliever, and saves became the standard measure of a relief pitcher’s success. There is an intuitive logic which says that if the best relief pitcher is the pitcher who gets the most saves, then the relief pitcher’s job is to “save” the game, and therefore the best use of a relief pitcher is to use him in save situations. It’s one of those things that seems to make sense unless you actually think about it, at which point you will realize that the caboose is not really hitched to the engine.

The trivial consequence of this change has been a tremendous growth in the number of saves acquired by the top relievers. It’s almost become a game for the managers, see how many saves you can get for your relief ace. Elroy Face in his career averaged less than one save every six innings of relief pitching. Dan Quisenberry established a major league record for saves in 1983, averaging less than one save every three innings. The top relievers today are closing in on one save per inning pitched. Quisenberry said, “I thought I was a stud. I thought I’d hold the record a few years, at least.” Little more than a decade later, his record 45 saves in a season has been pushed out of the top ten, and will soon enough be pushed out of the top 100.

But at the same time, the slow decay of the expectation that a starting pitcher will finish his games has proceeded. In 1980, starting pitchers still finished 20% of their games. This percentage had gone down in every ten-year period in baseball’s history:

Since 1980 it has continued to decline—indeed, the rate of descent has accelerated as the total has approached zero. In 1995, major league starting pitchers completed less than 7% of their starts. The trend line, analyzed without respect to other factors, would suggest that within twenty years there will be a major league season in which no pitcher throws a complete game.

So then we have two things which have happened in the last twenty years.

1) The number of innings pitched by relief aces has declined tremendously.

2) The number of complete games by starters has also declined.

So what does that mean? There has been a tremendous growth in the work assigned to middle relievers. In just fifteen years, there has been a 60% increase in the number of relievers used per game, as was shown in the chart at the beginning of this article.

Which means that roster spots which for sixty years were reserved for third catchers, sixth outfielders, and pure pinch hitters have been suddenly usurped by middle relievers. In 1965 teams normally carried nine pitchers, sometimes eight, and occasionally seven. Now twelve is standard, and sometimes we see thirteen. If you add three pitchers, you’ve got three less spaces for other players.

One thing you will hear, on almost any baseball broadcast, is a disparaging comment about the quality of pitching today, or about the
shortage
of good pitching nowadays, due to expansion. Now, in general I treat such comments with all the respect I might give to, let us say, the ethical guidance of Ted Kennedy. I can see absolutely no reason why expansion should create a shortage of pitching, as opposed to a shortage of hitting, and there is no evidence that any such thing has happened.

In every baseball game there is one win and one loss. Every hit for a hitter is a hit against a pitcher, and every out recorded by a pitcher is an out charged to some batter. Anything that happens on a baseball field can be presented either as a success or a failure. The people who insist on interpreting everything that happens as a failure tells us how terrible baseball is now and how much better it used to be are just miserable old farts who would do us all a favor if they would shut up and go home.

But let’s suppose, for a moment, that there
is
a shortage of pitching. Where would such a shortage come from? Since 1970, the number of major league teams has increased by one-sixth, or 17%. Our population has increased a little more than 20% in the same era, so that hardly explains why we would have a shortage of pitchers.

But in 1970, we had twenty-four teams carrying nine or ten pitchers each. Of the 849 players who played in the major leagues in 1970, 43% were pitchers. Of the 1,133 players who played in the major leagues in 1995, 49% were pitchers. The number of pitchers per team (sometime during the season) has increased from 15.1 to 19.6—while the number of nonpitchers has been essentially constant.

If there is, in fact, a shortage of pitching, couldn’t the rapid expansion of bullpens be a primary cause of it? Doesn’t that really make more sense, in calm review, than the popular theory that expansion has somehow diluted the quality of pitching, without any effect of the quality of hitting?

In 1965 the Chicago White Sox had an outstanding three-man bullpen:

Three pitchers, who among them pitched 400 innings. In 1995 the Chicago White Sox had a six-man bullpen, in which no one pitched more than 81 innings:

We have many more relief pitchers, making many more game appearances in the aggregate although somewhat fewer as individuals, but pitching dramatically fewer innings per outing, and with the save opportunities all given to a single reliever.

Which leaves us with two large questions:

1) Where are we going? and

2) Does where we are now make any sense?

For more than a hundred years, the number of complete games has been going down, and the number of pitchers per game has been going up. We’re now up to two and half pitchers per team per game, and accelerating rapidly. In another generation, will we see three or four or even five pitchers per game?

When a trend has been in motion for more than a hundred years, the forces behind it obviously have a great deal of power, and one should be cautious about predicting that the trend is about to turn around. Nonetheless, I believe that the 2.45 pitchers per game of 1995 is about as high as the average is likely to go, and that this number is as likely to go downward as it is to go upward. The standard bullpen of 2025 may well resemble the bullpen of 1965 more closely than it does the bullpen of 1995.

We will go backward, in my opinion, because:

1) it isn’t possible, in some respects, to go forward much further, and

2) the direction in which we have been going doesn’t make any sense.

Point a, we have gone, since 1890, from 89% complete games to 7%. We can’t go any lower than zero. We’ve gone about as far as we can go there.

Of course, we could go from starters being lifted in the seventh inning to starters being lifted in the sixth inning to starters being lifted in the fifth inning, so this may be a false argument, but, still on point a, the expansion of the size of the pitching staff has gone as far as it can go. The major league roster has been twenty-five men since before 1920. There is no evidence that the roster is about to expand; in fact, it is very clear that the rosters are
not
about to expand. You can’t carry sixteen pitchers and nine position players; that won’t work. We’ve got twelve-man staffs standard now; that’s about as far as we can go.

One place we
could
go, and it’s not too unlikely that we will, is toward more game appearances per pitcher. The bullpen of the future may consist of six men, but each of those six men may pitch 80 to 120 times per season, as opposed to the 50 to 80 which is standard now. A pitcher pitching 100 games a season, but facing two or three batters per outing, thus pitching only 50 to 70 innings in those 100 games … that probably is going to be a
part
of the future bullpen.

But there are two other elements of the 1990s bullpen that I feel strongly are going to break up, simply because the world is ultimately logical, and what major league managers are doing now
isn’t
logical. Those two elements are

1) The constant use of left-handed relievers to get out one or two hitters, and

2) The concentration of nearly all save opportunities onto a single reliever.

In 1994 Felipe Alou, manager of the Montreal Expos, did not have a left-hander in the bullpen. He had a five-man bullpen, with Tim Scott, Mel Rojas, John Wetteland, Gil Heredia, and Jeff Shaw. They were all good, but they were all right-handers.

Every time you saw the Expos on TV the announcer would tell you what a tremendous disadvantage Felipe Alou had, not having a left-hander in the bullpen. The Expos had the best record in baseball, 74–40, at the time the season was stopped by Bud Selig’s breath.

So I got to wondering, what does this really cost him, not having a left-hander out there? And I worked through the math, how many times does a normal manager have the platoon advantage when he brings in a pitcher, how often did Alou have that advantage, etc. How much does it cost him, per opportunity, and how many opportunities were involved?

And guess what? It doesn’t make any difference. Alou’s lack of a left-hander in the bullpen cost him the platoon advantage in less than 100 matchups over the course of a season. The normal platoon advatage is about .025, twenty-five points. So the lack of a left-hander in the bullpen probably cost him less than three hits.

We can suppose that the three hits are not all singles; hell, let’s suppose one of them is a home run. We can suppose that there might also have been a walk in there, or a situation where an intentional walk was forced by the lack of a lefty to come in. We can suppose that there might have been some hidden situations, situations I didn’t find, in which Alou also had a matchup disadvantage when another manager might have had an advantage, although frankly it’s hard to understand how. But you can add it up and throw in everything you can think of, and you can’t get to five runs.